© 2012 Macmillan Publishers Limited. All rights reserved.

LETTERSPUBLISHED ONLINE: 15 JULY 2012 | DOI: 10.1038/NCLIMATE1580

Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundingsTandong Yao1,2*, Lonnie Thompson1,3, Wei Yang1, Wusheng Yu1, Yang Gao1, Xuejun Guo1,Xiaoxin Yang1, Keqin Duan1,2, Huabiao Zhao1, Baiqing Xu1, Jiancheng Pu2, Anxin Lu1,2, Yang Xiang1,Dambaru B. Kattel1 and Daniel Joswiak1

The Tibetan Plateau and surroundings contain the largestnumber of glaciers outside the polar regions1. These glaciersare at the headwaters of many prominent Asian rivers andare largely experiencing shrinkage2, which affects the waterdischarge of large rivers such as the Indus3,4. The resultingpotential geohazards5,6 merit a comprehensive study of glacierstatus in the Tibetan Plateau and surroundings. Here wereport on the glacier status over the past 30 years byinvestigating the glacial retreat of 82 glaciers, area reductionof 7,090 glaciers and mass-balance change of 15 glaciers.Systematic differences in glacier status are apparent fromregion to region, with the most intensive shrinkage in theHimalayas (excluding the Karakorum) characterized by thegreatest reduction in glacial length and area and the mostnegative mass balance. The shrinkage generally decreasesfrom the Himalayas to the continental interior and is the leastin the eastern Pamir, characterized by the least glacial retreat,area reduction and positive mass balance. In addition to risingtemperature, decreased precipitation in the Himalayas andincreasing precipitation in the eastern Pamir accompanied bydifferent atmospheric circulation patterns is probably drivingthese systematic differences.

Although some glaciological studies have been done in theTibetan Plateau (TBP) and surroundings7–15, a region with a totalglacial area of ∼100,000 km2 (Supplementary Table S1), the recentcontroversies7,16,17 concerning glacial shrinkage in the Himalayasemphasize the necessity for a more comprehensive study. In addi-tion, more concrete in situ observation data will help to recheck theresults of a positive glacial mass balance of ∼7Gt yr−1 in Tibet andQilian Shan, which might be from uncertainty or misinterpretationof Gravity Recovery and Climate Experiment data7.

Under the progresses of the Third Pole Environmentprogramme18, an integrated assessment of glacier status in andaround the TBP over the past 30 years can now be provided. Datafor this assessment come from studying the glacial area reductionof 7,090 glaciers, with an area of approximately 13,363.5 km2 in the1970s and approximately 12,130.7 km2 in the 2000s (with a <5%uncertainty; see Supplementary Information) using topographicmaps and satellite images from Landsat-MSS/TM/ETM+, ASTERand LISS (Supplementary Tables S2 and S3 and Figs S1 and S2).Eighty-two glaciers were also studied for glacial retreat using insitu observations and previous studies (Supplementary Table S4)and 15 glaciers have undergone intensive study of glacial massbalance by in situ measurement (Supplementary Tables S5 andS6 and Figs S3–S15).

1Key Laboratory of Tibetan Environmental Changes and Land Surface Processes, Institute of Tibetan Plateau Research, Chinese Academy of Sciences,Beijing 100101, China, 2State Key Laboratory of Cryosphere Sciences, Chinese Academy of Sciences, Lanzhou 730000, China, 3Byrd Polar Research Centerand School of Earth Sciences, Ohio State University, Columbus, Ohio 43210, USA. *e-mail: tdyao@itpcas.ac.cn.

45° N

40° N

35° NLa

titud

e

30° N

70° E 75° E 80° E 85° E

Longitude

90° E 95° E 100° E

25° N East Asian monsoonIndian monsoonWesterliesELA (m a.s.l.)

3,0003,200

3,600 3,800

3,8004,0004,000

4,400

4,800

4,800

4,200

3,400

5,000

5,000

5,200

5,200

5,40

0

5,60

0

5,80

0

5,900

5,800

5,700

5,600

5,500

5,400 4,600

4,600

4,400

4,700

4,800

5,000

5,000

4,800

4,900

5,300

5,2005,100

5,000

Figure 1 | Distribution of glaciers and ELAs in and around the TBP11, whichare mainly under the dominance of the Indian monsoon and westerlies,with limited influence from the East Asian monsoon. Note the increasedglacier concentration and lower ELAs in the monsoon-dominatedsoutheastern TBP and the westerlies-dominated Pamir regions, comparedwith the sparse glacial distribution and high ELAs in thecontinental-climate-dominated interior.

Present atmospheric circulation patterns over the TBP andsurroundings are characterized by the Indian monsoon in thesummer and the westerlies in the winter (Fig. 1). These twocirculation systems, combinedwith the huge topographic landform,exert climatic controls on the distribution of existing glaciers.The East Asian monsoon also influences glaciers on the easternmargin, such as the Mingya Gongga and those in the easternQilian Mountains. The interior of the TBP is less influenced by theIndianmonsoon andwesterlies and dominatedmore by continentalclimatic conditions. As shown in Fig. 1, the high concentration andlow equilibrium line altitudes (ELAs) of glaciers in the southeasternTBP and the eastern Pamir regions result from high precipitationfrom the Indian monsoon and westerlies, respectively, whereasmore sparse glacier distribution and higher ELAs in the continental-climate-dominated interior are the consequences of limited water-vapour source from both these air masses.

To systematically and comprehensively assess glacier status inand around the TBP, we divided glaciers into seven regions,

NATURE CLIMATE CHANGE | VOL 2 | SEPTEMBER 2012 | www.nature.com/natureclimatechange 663

© 2012 Macmillan Publishers Limited. All rights reserved.

LETTERS NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE1580

75° E 80° E 85° E 90° E

Longitude Longitude

Longitude

95° E 100° E 75° E 80° E 85° E 90° E 95° E 100° E

75° E 80° E 85° E 90° E 95° E 100° E

40° N

35° N

30° N

Latit

ude

Latit

ude

Latit

ude

25° N

40° N

35° N

30° N

25° N

40° N

35° N

30° N

25° N

V

V

V

IV

IV

IV

VI

VI

VI

VII

VII

VII

¬0.9 ¬0.07

¬0.42

¬0.41

¬0.20

¬0.10

¬0.57

¬0.29

¬16.6 ¬16.3

II

II

II

III

III

III

¬2.0

¬4.2

¬6.3¬48.2

I

I

I

56.1¬80.6

28.2¬56.113.3¬28.24.10¬13.30¬4.10

0.54¬0.95

0.32¬0.540.17¬0.320.09¬0.170¬0.09

37.0¬82.0

16.3¬37.011.2¬16.34.40¬11.21.60¬4.40

Retreat (m yr¬1)

Advance (m yr¬1)

Muztag Ata

~ +250

~ ¬910

~ ¬550

~ ¬450

~ ¬400

~ ¬760~ ¬1,100

Hamtah Chhota ShigriNaimona'nyi

Area change rate (% yr¬1)

Xiaodongkemadi

Zhangxi

Gurenhekou

Parlung glaciers

Qiyi

YalaKangwure

AX 010Mass balance (mm w.e.)

0¬300300¬550550¬800800¬1,000

1,000¬1,700QiyiKangwureXiaodongkemadi

1975 1980 1985 1990 1995 2000 2005 2010

201020092008

Year

Year

20072006

2,0000

¬2,000¬4,000¬6,000¬8,000

¬10,000

Cum

ulat

ive

mas

s ba

lanc

e (m

m w

.e.)

Cum

ulat

ive

mas

s ba

lanc

e (m

m w

.e.)

2,0000

¬2,000¬4,000¬6,000¬8,000

¬10,000

Parlung No. 10Parlung No. 12Parlung No. 94Parlung No. 390

ZhongxiGurenhekou

KangwureNaimona'nyiMuztag AtaXiaodongkemadiQiyi

a b

d

e

c

Figure 2 | Spatial and temporal patterns of glacier status in the TBP and surroundings. a, Glacier retreat for 82 glaciers (Supplementary Table S4). b, Areareduction for 7,090 glaciers (Supplementary Tables S2 and S3 and Fig. S1). c, Mass balance for 15 glaciers (Supplementary Table S5). Glaciers arecategorized into seven regions and marked clockwise with Roman numerals in a–c. d, Cumulative mass balance for 11 glaciers in 2006–2010(Supplementary Table S6 and Figs S3–S13). e, Cumulative mass balance for the three longest time series of glacier mass-balance measurements alongtransect 1 (Supplementary Table S7 and Figs S14 and S15).

indicated clockwise as I–VII in Fig. 2. This classification allowsfor an ideal geographic representation of glaciers in the TBP andsurroundings, with one region each in the northeast and northwest,two in the centre and three in the Himalayas (southeastern TBP,central Himalayas and western Himalayas), which is the focus ofmany of the recent debates on glacier melt rates7,16,17. The sevenregions can also be categorized climatically into three transects:transect 1, southwest–northeast oriented (including regions III,II, VI and VII), with the weakening Indian monsoon influencenorthward; transect 2, southeast–northwest oriented (includingregions I, II and V), with the weakening Indian monsoon towardsthe interior and the strengthening westerlies towards the northwest;and transect 3, along theHimalayas (including regions I, III and IV),with stronger monsoon influence in the east and weaker monsooninfluence in the west.

Transect 3 (the Himalayas) shows the most extreme glacialshrinkage based on the reduction both of glacier length andarea (Fig. 2a,b and Supplementary Tables S2–S4). The shrinkageis most pronounced in the southeastern TBP (region I), wherethe length decreased at a rate of 48.2m yr−1 and the area wasreduced at a rate of 0.57% yr−1 during the 1970s–2000s. Therate of glacial shrinkage decreases from the southeastern TBP(region I) to the interior (regions II and VI). A decreasingtrend is discernible on both sides (regions III and VII) tothe interior (regions II and VI) along transect 1. The smallestrate of glacier contraction is observed in the eastern Pamirregions (region V), with a retreat rate of 0.9m yr−1 and area-reduction rate of 0.07% yr−1.

Mass balance is a direct and reliable indicator of glacier status.More than 20 glaciers have been measured for mass balance inthe TBP and surroundings, but only 15 have been measured overthree consecutive years (Supplementary Tables S5 and S6 andFigs S3–S13). Figure 2c shows the spatial pattern of the averageannual mass balance of these 15 glaciers in the seven regionsalong the three transects. Similar to the variations in glaciallength and area, the most negative mass balances occur along theHimalayas (transect 3), ranging from−1,100 to−760mmyr−1 withan average of −930mmyr−1. Along transect 2, the mass balanceis not so negative, ranging from −1,100mmyr−1 in region I to−400mmyr−1 in region II, then becoming positive in region VII(+250mmyr−1). Along transect 1, the least negative mass balanceis in the interior of region II.

Eleven of the fifteen glaciers have been continuously measured(from 2006 to 2010), thus enabling systematic studies by region(Fig. 2d; Supplementary Tables S5 and S6 and Figs S3–S13). Tenof the glaciers generally exhibit negative mass balance. The glacierwith the most negative mass balance is the Parlung No. 12 Glacier(29◦ 18′N, 96◦ 54′ E) in the Parlung River Basin in the southeasternTBP (region I), with an average of −1,698mmyr−1. In contrast tothe other ten glaciers, the Muztag Ata Glacier in the eastern Pamirregions (region V) had a positive mass balance in four of the pastfive observation years.

Based on mass-balance measurements of the Qiyi, Xiaodongke-madi and Kangwure glaciers, starting from 1975, 1989 and 1992respectively, and reconstructions along transect 1 (Fig. 2e; Supple-mentary Table S7 and Figs S14 and S15), we present an extended

664 NATURE CLIMATE CHANGE | VOL 2 | SEPTEMBER 2012 | www.nature.com/natureclimatechange

© 2012 Macmillan Publishers Limited. All rights reserved.

NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE1580 LETTERS

Qiangyong

Parlung 12

0 250 500km

Ata

Halong

Halong 1981 Ata 1933

Ata 1976

Ata 2006

Halong 2006

Qiangyong 1975

Qiangyong 2006

3,0002,5002,0001,5001,0005000

Retreated length since 1917 (m

)

Retr

eat r

ate

(m y

r¬1 )

020406080

100

5,10

0

5,200

5,400

5,300

5,200

5,10

0

5,200

5,400

5,300

5,2000 250

m500

1980 2005

5,100

Lake

N

f

e

d

Contour (m)

Glacier

a

g

W E

N

S

b

1915 1925 1935 1945 1955 1965Year

1975 1985 1995 2005

1980

19171933

19541973

1976

19801985

1990

19951999

20042005

AtaDongxin

Ata

Gla

cier

Ata Q

u

2,60

02,

800

3,0

00

3,20

03,

40

03,

600

3,80

04

,00

04

,20

0

3,000

1999

N

0 1 km

29° 6.20' N

29° 1.63' N

96° 4

5.71

' E

2005

96° 5

0.3

3' E

c

Contour (m)

Seasonal route

River

Glacier contour line and age

Glacier

Rock

Forest

Village

Figure 3 | Case studies of glacial retreat. a, The locations of studied glaciers. b, Area reduction of the Parlung No. 12 Glacier. c,d, Variations of the AtaGlacier since 1917. e–g, Glacial variation from comparison of photos of the Halong, Qiangyong and Ata glaciers, respectively. Panel e: top image © 1981courtesy of Matthias Kuhle; bottom image © 2005 Greenpeace/Dan Novis; Panel g: top image reproduced with permission from ref. 21, © 1933 Wiley.

history. Mass balances of the Qiyi and Xiaodongkemadi glacierswere positive until the early 1990s, then turned increasingly neg-ative over time. The values associated with the Kangwure Glacierhave been increasingly negative since 1992. In general, all theseglaciers show an accelerating trend of more negative mass balancesince the early 1990s.

Several glaciers in the TBP and surroundings were targeted forcase studies (Fig. 3). The photos in Fig. 3b,d clearly demonstrateintensive glacial shrinkage. Photographs taken in 1981 and 2006of the Halong Glacier (34◦ 46′N, 99◦ 30′ E, Fig. 3b; ref. 19) reveala significant reduction in length. In situ observations indicate thatthe retreat averaged 10m yr−1 during those 25 years. Photos of theQiangyong Glacier (29◦ 57′N, 90 ◦ 09′ E), which were published in1976 (ref. 20), are compared with our 2006 photos (Fig. 3c). Fromthese we can ascertain that the length has retreated at an average rateof 4m yr−1 during this period.

The Ata Glacier (29◦ 10′N, 96◦ 48′ E), with a length of 16.7 kmand an area of 13.8 km2, is on the southern slope of MountKangri Karpo in the southeastern TBP (Fig. 3d). The glacier wasphotographed in 1933 (ref. 21), 1976 (ref. 20) and 2006 (this study).We studied this glacier in detail using the historical photos, in situ

observations and the 1917 1:125,000 topographic map compiledby the US Army (www.lib.utexas.edu/maps/ams/china; Fig. 3f,g).The glacial reduction averaged less than 20m yr−1 between 1917and 1954, increased to more than 30m yr−1 from 1955 to 1984,increased again to more than 60m yr−1 after 1985 and exceeded80m yr−1 in 2005/2006. The history of the Ata Glacier thusillustrates the largest amplitude of glacial retreat and reveals anaccelerating trend in retreat since the 1990s.

A comparative study was done on the Parlung No. 12 Glacier,which has a length of 0.6 km and an area of 0.21 km2 (2005data) and is located on the northern slope of Mount KangriKarpo (Fig. 3a). From 1980 to 2005, it retreated by ∼700m andits area reduced by ∼55.3% (Fig. 3e). During field observationsfrom 2006 to 2010, the glacier thinned by 9.5m (21.1%) basedon measuring stakes and its ice volume reduced by 2.0× 106 m3

(38.7%) based on ground-penetrating radar measurements. If thisrate continues, the Parlung No. 12 Glacier will disappear withineither 25 years (based on ice-volume-reduction rates) or 35 years(based on ice-thickness-thinning rates). In contrast, the Ata Glacierwill not disappear in the near future given its larger size and overalllength. This implies a large-scale glacial behaviour with global

NATURE CLIMATE CHANGE | VOL 2 | SEPTEMBER 2012 | www.nature.com/natureclimatechange 665

© 2012 Macmillan Publishers Limited. All rights reserved.

LETTERS NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE1580

1970 1980 1990Year

2000 2010

1,000

900

800

700

600

Rain

fall

(mm

)

2

1

0

¬1

¬2

¬3Indi

an m

onso

on in

dex

Zon

al in

dex

1970 1980 1990Year

2000 2010

2

1

0

¬1

¬2

Standardized precipitation

Precipitation (mm

)

Prec

ipita

tion

(mm

)

200

220

240

260

280

300

3.0

1.5

0.0

¬1.5

¬3.0

1 2

3 4

5 6

7 8 9

10 11 12 13 14

15 16

40° N

35° N

Latit

ude

30° N

25° N95° E75° E 80° E 85° E

Longitude

Precipitation variation (mm d¬1)

90° E 100° E

¬0.07 ¬0.06 ¬0.05 ¬0.04 ¬0.03 ¬0.02 ¬0.01 0.00 0.005 0.010 0.015 0.020

b

c

d

e

f

g

a

0

40

80

120

160

Figure 4 | Precipitation patterns in the Himalayas and eastern Pamir regions. a, Precipitation trends for 1979–2010 from the GPCP. Plus symbols indicategrid centres (Supplementary Figs S17 and S18). b, All-India rainfall (June–September). c, Standardized June–September precipitation from Linzhi, Zayu andBomi in the southeastern TBP. d, Indian monsoon index for June–September30. e, Annual precipitation in Pamir regions and central Asia. f, Annualprecipitation at Taxkorgen in the eastern Pamir regions. g, Zonal index for the Northern Hemisphere29. Time series in b–g cover the period 1970–2009.Straight lines indicate the linear trends.

warming such that smaller glaciers retreat faster than larger ones(Supplementary Fig. S2).

Temperature rise22,23 contributes to glacial shrinkage over theTBP. Other factors are also important for glacier status, includingsize, type (temperate, polythermal or cold), condensation levels,sublimation, topography and debris cover, or a combination ofmany factors14,24. Climatic differences between low and high alti-tudes involve important issues relating to warming. An increasingwarming trend at higher elevations has been observed over theTBP (ref. 22) and the warming rate increases with elevationbefore becoming quite stable with a slight decline near the highestelevations23. The rate of warming is highest between 4,800 and6,200m above sea level (a.s.l.) (ref. 23), which includes the ablationaltitudes of almost all glaciers on the TBP. However, this is stillnot conclusive and more observations are necessary. Glaciers aremostly temperate in the southeastern TBP and on the southernslope of the Himalayas, whereas in other regions mentioned earlierthe glaciers are primarily cold. This may also contribute to theobserved differences in glacier status. However, the systematicdifferences suggest a precipitation pattern driven by atmospheric

circulation. Figure 4a shows the large-scale pattern of precipita-tion from the Global Precipitation Climatology Project (GPCP,www.esrl.noaa.gov/psd/data/gridded/data.gpcp.html) over the TBPand surroundings. There is strong evidence that precipitation from1979 to 2010 decreased in the Himalayas and increased in theeastern Pamir regions. The all-India rainfall (June–September;www.tropmet.res.in; Fig. 4b) and standardized June–Septemberprecipitation at three stations Linzhi (2,992m a.s.l.), Zayu (2,328ma.s.l.) and Bomi (2,736m a.s.l.) in the southeastern TBP (Fig. 4c)confirm the decreasing precipitation in the Himalayas from theGPCP. The annual precipitation in the eastern Pamir regions andcentral Asia (www.cru.uea.ac.uk; Fig. 4e) and at the Taxkorganstation (Fig. 4f) also confirms the increasing precipitation trendin the eastern Pamir based on the GPCP data. Recent studiesfound that the Indian monsoon is weakening25,26 and the westerliesare strengthening27, which can influence changes in precipitationpatterns28. The decreasing trend of the Indian monsoon index(Fig. 4d) suggests that the precipitation decrease in the Himalayasmay be linked to the weakening Indian monsoon, whereas theincreasing zonal index29 (Fig. 4) reveals that the precipitation

666 NATURE CLIMATE CHANGE | VOL 2 | SEPTEMBER 2012 | www.nature.com/natureclimatechange

© 2012 Macmillan Publishers Limited. All rights reserved.

NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE1580 LETTERSincrease in the eastern Pamir is linked to the strengtheningwesterlies. The general patterns ofmass balance over the TBP followatmospheric circulation patterns (Supplementary Fig. S16).

The glacier status in the TBP and surroundings varies system-atically from region to region: the Himalayas shows the greatestdecrease in length and area, and the most negative mass balance,whereas the eastern Pamir shows the least reduction in lengthand area, and positive mass balance. The main cause for thisregional trend is probably decreasing/increasing precipitation inthe Himalayas/eastern Pamir regions, which results from changesin the two different atmospheric circulation patterns, that is, theweakening Indian monsoon and strengthened westerlies. Underthe present warming conditions, glacier shrinkage might furtheraccelerate in the Himalayas whereas glaciers might advance in theeastern Pamir regions. Potential consequences of glacier changeswould be unsustainable water supplies from major rivers3,4 andgeohazards (glacier-lake expansion, glacier-lake outbursts andflooding)5,6, which might threaten the livelihoods and wellbeingof those in the downstream regions.

MethodsMass-balance measurement and calculation. Mass balance, specific net ablationand net accumulation were calculated frommeasurements in the field. Net-ablationmeasurements were carried out using the measuring-stake method in the ablationzone. Net-accumulation measurements were carried out using snow-pitmeasurements in the accumulation zone. In the measuring-stake method, snow-and ice-surface changes caused by negative net balance were manually determinedon measuring stakes. Snow-pit measurements involved mapping of visiblestratigraphic features. For both methods, the measurements were made at the endof each ablation season (generally at the end of September or beginning of October)and snowdensities were alsomeasured forwater-equivalent (w.e.) calculation.

For a given glacier, the overall glaciermass balanceB is calculated as

B=∑n

i=1bisiS

(in mm)

where bi is the specific mass balance (net ablation or net accumulation) of thegiven altitudinal range i over map area si and S is the total glacial area. For agiven altitudinal range, bi is obtained from the corresponding net-ablation ornet-accumulation measurements.

Glacial length observation. Annual variations of glacial length were observedand calculated by repeated observations between the benchmark locations andglacier termini. The uncertainty of present field observation by differential globalpositioning systems is negligible. The uncertainty of the previous field observationsis determined by the number of in situ measurements. Our measurements in thefield comprise five points for each small glacier (<2 km2) and nine points for largerglaciers. The uncertainty of this method is estimated at 5–10%.

Glacial area analysis. Among the glacial area analysis of 16 river basins in sevenregions, nine are gleaned from the literature and seven are based on our ownstudies (four have not been published). Topographic maps, aerial photography anddata from Hexagon KH-9, LISS-III/LISS-IV, Landsat MSS, Landsat TM/ETM+,ALOS AVNIR-2, Terra ASTER and SRTM DEM were considered in this study.As topographic maps, aerial photography and remote sensing data were taken atdifferent times and different resolutions, they were first orthorectified, co-registeredand correlated. The TM3/TM5, TM4/TM5 band-ratio methods were used toautomatically delineate the glacial area in our study. After automated delineation,we visually checked and manually adjusted the regions for shadow, seasonalsnow, turbid/frozen/multihued proglacial lakes and debris cover. The mappinguncertainty of our studies is less than 3% for clean-ice glaciers and 4% fordebris-covered glaciers. The methods and results from previous studies includemanual delineation based on visual interpretation from digitized topographic mapsand/or false-colour composite satellite images, supervised classification, band-ratiomethod, normalized different snow index and normalized different water index.The uncertainty of those studies is ±2–3% for clean-ice glaciers and ±3–4% fordebris-covered glaciers for ASTER and Landsat TM (see the third paragraph in theSupplementary Information).

Received 21 November 2011; accepted 14 May 2012;published online 15 July 2012

References1. Yao, T. et al. Map of Glaciers and Lakes on the Tibetan Plateau and the

Surroundings (Xi’an Cartographic Publishing House, 2008).

2. Yao, T., Pu, J., Lu, A.,Wang, Y. & Yu,W. Recent glacial retreat and its impact onhydrological processes on the Tibetan Plateau, China, and surrounding regions.Arct. Antarct. Alp. Res. 39, 642–650 (2007).

3. Immerzeel, W.W., Beek, L. P. H. & Bierkens, M. F. P. Climate change will affectthe Asian water towers. Science 328, 1382–1385 (2010).

4. Kaser, G., Großhauser, M. & Marzeion, B. Contribution potential of glaciersto water availability in different climate regimes. Proc. Natl Acad. Sci. USA 107,20223–20227 (2010).

5. Richardson, H. D. & Reynolds, J. M. An overview of glacial hazards in theHimalayas. Quat. Int. 65/66, 31–47 (2000).

6. Kääb, A. et al. Remote sensing of glacier- and permafrost-relatedhazards in high mountains: An overview. Nat. Hazards Earth Syst. Sci. 5,527–554 (2005).

7. Jacob, T., Wahr, J., Pfeffer, W. T. & Swenson, S. Recent contributions ofglaciers and ice caps to sea level rise. Nature 482, 514–518 (2012).

8. Ding, Y., Liu, S., Li, J. & Shangguan, D. The retreat of glaciers in response torecent climate warming in western China. Ann. Glaciol. 43, 97–105 (2006).

9. Li, X. et al. Cryospheric change in China. Glob. Planet. Change 62,210–218 (2008).

10. Bolch, T. et al. The state and fate of Himalayan glaciers. Science 336,310–314 (2012).

11. Shi, Y., Liu, C. & Kang, E. The glacier inventory of China. Ann. Glaciol. 50,1–4 (2009).

12. Li, Z., Sun, W. & Zeng, Q. Measurements of glacier variation in the TibetanPlateau using Landsat data. Remote Sens. Environ. 63, 258–64 (1998).

13. Fujita, K., Ageta, Y., Pu, J. & Yao, T. Mass balance of Xiao Dongkemadiglacier on the central Tibetan Plateau from 1989 to 1995. Ann. Glaciol. 31,159–163 (2000).

14. Scherler, D., Bookhagen, B. & Strecker, M. R. Spatially variable response ofHimalayan glaciers to climate change affected by debris cover. Nature Geosci.4, 156–159 (2011).

15. Fujita, K. &Nuimura, T. Spatially heterogeneous wastage ofHimalayan glaciers.Proc. Natl Acad. Sci. USA 108, 14011–14014 (2011).

16. Cogley, J., Kargel, J. S., Kaser, G. & van der Veen, C. Tracking the source ofglacier misinformation. Science 327, 522–522 (2010).

17. Bagla, P. No sign yet of Himalayan meltdown, Indian report finds. Science 326,924–925 (2009).

18. Yao, T. & Greenwood, G. A new polar program. EOS 90, 515–515 (2009).19. Greenpeace & Kuhle, M. Yellow River source at risk under climate change.

(http://activism.greenpeace.org/yellowriver/yrs-english_web.pdf, 2007).20. Li, J. et al. Glaciers in Tibet (in Chinese) (Science Press, 1986).21. Ward, F. K. The Himalaya East of the Tsangpo: A paper read at the

evening meeting of the society on 30 April 1934. Geograph. J. 84,369–397 (1933).

22. Liu, X. & Chen, B. Climatic warming in the Tibetan Plateau during recentdecades. Int. J. Climatol. 20, 1729–1742 (2000).

23. Qin, J., Yang, k., Liang, S. & Guo, X. The altitudinal dependence of recent rapidwarming over the Tibetan Plateau. Climatic Change 97, 321–327 (2009).

24. Xu, B. et al. Deposition of anthropogenic aerosols in a southeastern Tibetanglacier. J. Geophys. Res. 114, D17209 (2009).

25. Wu, B. Weakening of Indian summer monsoon in recent decades.Adv. Atmos. Sci. 22, 21–29 (2005).

26. Thompson, L.G. et al. Abrupt tropical climate change: Past and present.Proc. Natl Acad. Sci. USA 103, 10536–10543 (2006).

27. Zhao, H. et al. Deuterium excess record in a southern Tibetan ice core and itspotential climatic implications. Clim. Dynam. 38, 1791–1803 (2012).

28. Naidu, C.V. et al. Is summer monsoon rainfall decreasing over India in theglobal warming era? J. Geophys. Res. 114, D24108 (2009).

29. Li, J. & Wang, J. X. L. A modified zonal index and its physical sense.Geophys. Res. Lett. 30, 1632 (2003).

30. Wang, B., Wu, R. & Lau, K-M. Interannual variability of the Asian summermonsoon: Contrasts between the Indian and the western North Pacific-EastAsian monsoons. J. Clim. 14, 4073–4090 (2001).

AcknowledgementsThis work is supported by the NSFC (41190081, 40810019001), the CAS (ExternalCooperation Program GJHZ0960 and SAFEA International Partnership Programfor Creative Research Teams) and the MOST (2005CB422004). We thank Q. Ye forhelp preparing Fig. 3.

Author contributionsAll authors contributed extensively to this work.

Additional informationThe authors declare no competing financial interests. Supplementary informationaccompanies this paper on www.nature.com/natureclimatechange. Reprints andpermissions information is available online at www.nature.com/reprints. Correspondenceand requests for materials should be addressed to T.Y.

NATURE CLIMATE CHANGE | VOL 2 | SEPTEMBER 2012 | www.nature.com/natureclimatechange 667